|Bag of Sampled Words: A Sampling-based Strategy for Fast and Accurate Visual Place Recognition in Changing Environments
Sang Jun Lee and Sung Soo Hwang*
International Journal of Control, Automation, and Systems, vol. 17, no. 10, pp.2597-2609, 2019
Abstract : In the field of visual place recognition, a variety of methods using Visual Bag of Words has been suggested to cope with environmental change. This paper presents a sampling-based method which improves the speed and the accuracy of the existing Visual Bag of Words models. We first propose sampling of image features considering their density to speed up the quantization step. By using samples, a more accurate but rather slow ranking procedure is feasible. Thus, we also propose a ranking procedure which utilizes spatial information of samples. Lastly, a coarse and fine approach-based refinement method is proposed which increases the accuracy of the system by iteratively updating the similarity between images. The experimental results show that the proposed method improves the performance of the existing Visual Bag of Words models in terms of speed and accuracy.
Bag of words, image retrieval, visual place recognition.
Download PDF : Click this link